Tuesday 12 March 2019

Horizon's robustness

A Boeing 787 Max 8
It is apparent the Post Office won the early negotiations on which the Horizon trial's battleground was decided. That is conceded by the claimants in their opening argument. They say they did not want the "robustness" of Horizon to be part of the trial (see day 1 transcript). It is.

In fact, the Post Office insists it is central to the trial. Which led to Anthony de Garr Robinson, the Post Office QC saying yesterday:
"In short, it is accepted that Horizon works well for the overwhelming majority of cases and consistently with that it is now common ground between the experts that Horizon is robust and that its robustness has improved over time and your Lordship already has the reference, it is the joint statement, the third joint statement…  
Now, what does relatively robust mean?  It means robust as compared with comparable systems big systems, systems that keep aircraft in the air, that run power stations and that run banks."
This statement is essentially an admission that Horizon works well for most people most of the time, except when it occasionally catastrophically fails.

Today, in the immediate aftermath of a Boeing 737 crash which killed 157 people, Boeing have announced a software update to the 737 Max 8 fleet. Boeing are, of course, not linking the update with what happened to the Ethiopian Airways plane at Addis Ababa, but surely no one would be surprised if software, rather than pilot error was to blame. Both computers and planes are operated by humans.

Tony Collins is a former editor of Computer Weekly. His pioneering work into possible software errors in the Chinook helicopter crash on the Mull of Kintyre helped overturn the verdict of pilot error. Today he has republished a piece on his blog which describes the conclusion of a National Transportation Safety Board investigation into the USAir Flight 427 Boeing 737 crash in 1994, which killed 132 people. Boeing had, for years, blamed the pilots. Turned out "it was an intermittent malfunction – and one that occurred in a rare set of circumstances. It left no trace. It might have caused a succession of seemingly-unique major incidents."

Yesterday Mr de Garr Robinson said the claimants' propensity to blame Horizon for branch accounting errors "is driven by the natural human scepticism to technology."

This fits neatly into a historic pattern of big companies like Boeing blaming operator error until after several long court cases and at great expense, a different picture emerges. It might be that the Post Office will be proved correct on the facts, but both the assertions I have posted above suggest Mr de Garr Robinson hasn't done his homework.